Ffrith, Flintshire, Wales Archaeological Evaluation & Assessment of the Results

Wessex Archaeology

Ref: 59461.01 August 2005 FFRITH, FLINTSHIRE, WALES

ARCHAEOLOGICAL EVALUATION AND ASSESSMENT OF THE RESULTS

Prepared for:

Videotext Communications 49 Goldhawk Road LONDON SW1 8QP

by Wessex Archaeology

Report reference: 59461.01

August 2005

ARCHAEOLOGICAL EVALUATION AND ASSESSMENT OF THE RESULTS

Contents

Summary Acknowledgements

1 Introduction ...... 1 1.1 Project Background...... 1 1.2 Archaeological Background...... 1 2 Aims and Objectives...... 3 2.1 Introduction...... 3 2.2 Research Questions and Aims...... 3

3 METHODOLOGY...... 3 3.1 Survey ...... 3 3.2 Geophysical survey...... 4 3.3 Strategy ...... 4 3.4 Excavation and Recording...... 4 3.5 Finds...... 4 3.6 Environmental and Scientific Sampling ...... 4

4 RESULTS...... 5 4.1 Introduction...... 5 4.2 Geophysical Survey...... 5 4.3 Possible Timber Buildings...... 5 4.4 Masonry Structures...... 6 4.5 Other Features and Deposits ...... 7

5 Finds...... 9 5.1 Introduction...... 9 5.2 Pottery ...... 9 5.3 Ceramic Building Material...... 10 5.4 Opus Signinum, Mortar and Wall Plaster ...... 10 5.5 Fired Clay...... 10 5.6 Clay Pipe...... 11 5.7 Stone...... 11 5.8 Glass...... 11 5.9 Slag...... 11 5.10 Metalwork ...... 11

i 5.11 Worked Bone...... 11 5.12 Animal Bone...... 12

6 Palaeoenvironmental evidence...... 12

7 Discussion ...... 14

8 Recommendations...... 15

9 Archive...... 16

10 References ...... 17

Appendix 1: Catalogue of Trench Descriptions

Table 1: Finds Totals by Trench

Table 2: Pottery Totals by Ware Type

Table 3. Assessment of the charred plant remains and charcoal

Figure 1. Location plan showing trenches, areas of previous work and scheduled areas Figure 2. Location plan showing geophysical survey areas and interpretation Figure 3. Trenches 1 and 6 (Phases 1 and 2) Figure 4. Trenches 1, 2, 4, 6 and 9 (Phase 3) and previously investigated areas

ii SUMMARY

Wessex Archaeology was commissioned by Videotext Communications Ltd to carry out archaeological recording and post-excavation analysis on an archaeological evaluation by Channel 4’s ‘Time Team’ at Ffrith, Flintshire, Wales (NGR 328500 355300). The fieldwork, comprising one hand-excavated and nine machine-excavated trenches, was undertaken between 19th and 21st April 2005.

A considerable number of Romano-British artefacts have been found throughout the village, many in association with possible building remains, one of which had been previously investigated. Although two areas within the village have been designated as Scheduled Monuments (SAM 164a and 164b), the nature of the Romano-British remains is not clearly understood. It is possible that they form part of a small complex of buildings or they may represent part of a larger and more extensive settlement.

It was hoped that a project combining geophysical survey with trial trenching would clarify a number of key issues about the nature of the Romano-British remains at Ffrith. It was proposed that during the course of this project, two main areas were to be investigated by excavation and survey. The playing fields (SAM 164a) are situated in a natural, flat promontory formed by the confluence of two streams and any Romano-British activity may have utilised this space. The car park of Blue Bell pub and the garden of ‘The Glen’ next to it (SAM 164b) are very close to the northern stream and a previously investigated Romano-British building. In addition to these two main areas, the project also included trenching in the gardens of houses near to the site of the previously excavated remains.

The results from both the magnetic and resistance geophysical surveys in the playing field (SAM 164a) proved disappointing. The magnetic data were severely distorted by both modern and earlier playground furniture. In addition, metal fences and buildings on the periphery of the survey grids resulted in disturbed areas. The resistance survey identified a few regions of high resistance; however, these proved to be of natural origin. Ground penetrating radar survey identified Victorian and modern features in the Blue Bell car park and adjacent garden, but no Romano-British features or deposits. Resistance survey in one back garden proved inconclusive as did a GPR survey of a driveway / parking area.

Excavation in the area of previous excavations exposed three of the previously identified masonry walls, all of which appeared to be broadly contemporaneous, overlying the truncated remains of earlier, possible timber, buildings. Only a very small area of the possible timber building or buildings, comprising small areas of clay floors, several stake holes, a hearth and associated occupation deposits were investigated. Although finds from these deposits were scarce, and any dating is therefore tentative, an early 2nd century date is suggested. Environmental evidence from these deposits, along with the few finds and the structural remains themselves, suggest a domestic function.

The three masonry walls were dated to the 2nd or early 3rd century. The walls were all constructed using unworked, or roughly trimmed, local sandstone and limestone; however, three different building methods were apparent. A well-built mortar bonded wall with associated mortar floor and external surface was interpreted as part of a

iii substantial building, which earlier observations suggested contained at least one room with a hypocaust. Approximately 7m to the south was a clay bonded wall of broadly similar date. This appeared to represent the north-western corned of a separate structure, possibly a small workshop or storage building. The third wall, a curvilinear dry stone wall, was traced for approximately 8m. Earlier excavations suggested that this represented the apsoidal end of a substantial building. However, the rather crude construction and irregular form, along with the complete absence of associated floors or surfaces indicate that this is very unlikely and is more probably an enclosure wall, possibly surrounding the substantial building to the north and east. The truncated remains of a 2nd century ditch was found approximately 12m to the north-west of the mortared wall and is assumed to be broadly contemporary.

A possible buried topsoil and an overlying alluvial deposit, both of which contained Romano-British material were located in trenches at the southern end of the village, but no features of this date were found. All of the Romano-British features and deposits encountered during the evaluation were concentrated in the area of the earlier excavations. Trenches elsewhere in the village, while often producing residual Romano-British finds, were generally devoid of features and deposits of this date. The results of the evaluation therefore suggest that the Romano-British settlement at Ffrith comprised at least one substantial and relatively high status building, possibly with other ancillary buildings, surrounded by a dry stone wall. However, this appears to have been confined to the northern end of the present village.

iv ACKNOWLEDGEMENTS

The evaluation was commissioned and funded by Videotext Communications Ltd. Wessex Archaeology is grateful to the landowners, Harry and Tracy Leese (Arosfa), Gwen Jones (Edlyingar), David Stone (Blue Bell), Karl Arndt (The Glen), Carolyn Thomas (Clerk for Playing Fields) and Bill Roberts (parking area by garages), for their assistance during the course of the fieldwork. Thanks are also due to Chris Martin, Head of Curatorial Services at CPAT and Sian Rees of CADW.

Surveying was undertaken by Henry Chapman, University of Hull, and the geophysical survey was undertaken by John Gater, Jimmy Adcock and Fiona Robertson of GSB Prospection. The excavation was undertaken by Phil Harding, Kerry Ely, Matt Williams, Raksha Dave, Brigid Gallagher and Ian Powlesland of Time Team and Fiona Grant, Ian Davis, George Luke, Dan Garner and Lee Dodd of Clwyd-Powys Archaeological Trust, who were assisted by local metal detectorists Bill Smuts and Paul Stanway.

The archive was collated and all post-excavation assessment and analysis undertaken by Wessex Archaeology, including management (Roland Smith), finds (Stephanie Knight, Lorraine Mepham), environmental assessment (Sarah F Wyles, Michael J Allen, Cathie Chisham and Chris Stevens), report (Vaughan Birbeck), and illustrations (Matthew McMurray).

v FFRITH, FLINTSHIRE, WALES

ARCHAEOLOGICAL EVALUATION AND ASSESSMENT OF THE RESULTS

1 INTRODUCTION

1.1 Project Background 1.1.1 Wessex Archaeology was commissioned by Videotext Communications Ltd to carry out archaeological recording and post-excavation analysis on an archaeological evaluation undertaken by Channel 4’s ‘Time Team’ at Ffrith, Flintshire, Wales (centred on NGR 3285 3553). The fieldwork, comprising one hand-excavated and nine machine-excavated evaluation trenches was undertaken between 19th - 21st April 2005 by Time Team and local archaeologists.

1.1.2 The evaluation area comprised the village of Ffrith, in Flintshire, Wales (Fig 1). Ffrith is 20km southwest of Chester and 7km northwest of Wrexham on the Welsh border. Ffrith is situated within a wooded valley at the confluence of two rivers: the Nant–y–Ffrith and the Cegidog and lies at between 120m and 124m OD. The underlying geology comprises sandstone overlain by alluvial deposits. The village is largely of 20th century date with only a few buildings dating to the 19th century; the village church was constructed in 1842 and by 1850 only a half dozen houses existed, all south of the river Cegidog.

1.2 Archaeological Background 1.2.1 The earliest traces of human activity noted in the area comprised an assemblage of Mesolithic flintwork recovered from probable alluvial deposits in the 1960’s (Blockley 1989, 143). However, the main purpose of the Time Team evaluation was to investigate the character, extent and date of later Romano-British remains. Earlier archaeological investigations have established the presence of Romano-British remains below the village, however, the extent and character of these remains uncertain.

1.2.2 Ffrith lies on the Roman road from Chester (Deva) to Caer Gai, some 21 km from Chester, in an area which was widely occupied during the Roman period. The military presence at nearby Chester is believed to have influenced the activity at Ffrith, although it is not understood if the remains here represent military or civilian occupation. Chester was the home of the 20th Legion during the Roman period, whose works depot was based at Holt (Bovium), some 6km to the east of Ffrith. Two areas of the village are designated Scheduled Ancient Monuments (SAM 164a and b), the first encompasses the village playing fields; the second is the car park of the Blue Bell public house and an adjacent garden (the Glen). Their proximity to earlier excavations and observations suggest that these undeveloped areas are likely to yield more evidence of Romano-British occupation.

1 1.2.3 The Romano-British remains are likely to have been effected by the construction of Offa’s Dyke. This 8th century monument, built by the King of Mercia, runs approximately north-south along the length of the Main Street and survives as a low, linear earthwork clearly visible in the playing fields, to the east of the High Street. The main purpose of the Time Team evaluation was, however, to investigate the character, extent and date of the Romano- British settlement.

1.2.4 The presence of the Romano-British remains was noted as early as c. 1586, when the site was described in Camden’s Britannia. Camden described a hypocaust system with brick vents and tiles stamped with LEG XX, although no definite find spot is recorded. In 1828 a part of Offa’s Dyke was levelled and Roman coins, jewellery, pottery and probable cremations were found along with a “votive altar with a mutilated inscription” (Davies 1949, 230-1), once again, no definite find spot is recorded. The remains of a hypocaust were also recorded in 1870 and 1874 “when cutting the foundations of some new houses in the angle formed by the High Road at Ffrith and that leading to the Blue Bell” (Davies 1949, 231). In 1910 further Romano-British remains, comprising masonry walls, were exposed during the construction of a new house (Arosfa) on the western side of the High Street (Davies 1949, 232).

1.2.5 The archaeological excavation of a trench across Offa’s Dyke was undertaken in 1926 by Sir Cyril Fox in the southern end of the village and a small assemblage of Romano-British pottery of probable 2nd century date was recovered from below the Dyke. The further levelling of the Dyke in 1933, to create an access into the recreation ground, also produced Romano-British finds, although these do not appear to have been kept (Davies 1949, 233).

1.2.6 In the spring of 1967, a group of amateur archaeologists begin some small- scale excavations in the gardens of private houses near to Arosfa. A curvilinear sandstone wall was traced for approximately 60 feet (18m). Further masonry walls and associated and earlier floors were also located to the north and east. Artefacts associated with the walls included animal bones, samian ware pottery and a tile bearing the stamp ‘LEG XX’. The walls were interpreted at the time of excavation as part of a substantial building with an apse-ended room, possibly part of a bath house. Contemporary photographs show substantial walling and some undisturbed deposits of Romano-British material, opus signinum floors and wall plaster. The possible remains of a timber building, or buildings were also recognised below the masonry remains (Blockley 1989, 140), which were left in situ.

1.2.7 The most recent archaeological work took place in 1990, when Clwyd-Powys Archaeological Trust (CPAT) carried out a rescue excavation during the construction of a new house adjacent to the western gate of the playing fields, near to the site of Cyril Fox’s dig. Roman pottery dating from the 1st to mid 2nd century was recovered from the old ground surface sealed beneath Offa’s Dyke. A hard stony surface was also recorded, and it is speculated that this may be the metalling of the roman road through the village, which now lies beneath the dyke.

2 2 AIMS AND OBJECTIVES

2.1 Introduction 2.1.1 A project design for the work was compiled by Videotext Communications (Videotext Communications 2005), providing full details of the circumstances and methods of the project, as summarised here.

2.1.2 The project offered the opportunity to evaluate and assess the character of the Romano-British remains in Ffrith. It was hoped that a project combining geophysical survey with trial trenching would clarify a number of key issues about the nature of the Romano-British remains at Ffrith. It was proposed that during the course of this project, the two Scheduled Areas were to be investigated by excavation and survey. In addition to these two main areas, the project also included investigations in the gardens of houses near to the 1967 dig site.

2.2 Research questions and aims 2.2.1 This project attempted to determine the nature of the Romano-British remains at Ffrith, by considering a series of fundamental questions about the origins, date and longevity of the site. The general questions identified in the research design were: x What was the function of the Romano-British building partly excavated in 1967, and is it an isolated structure? x What is the nature of the Romano-British activity at Ffrith? Is it civilian, military or industrial? x Is there any relationship with the Romano-British military settlements at Chester and Holt? x Are the artefacts discovered during 1926 (Fox) and 1990 (CPAT) associated with Romano-British buildings or settlement, and if so are they contemporary with the building excavated in the gardens of Arosfa and surrounding houses? x What is the nature and extent of the settlement? x When was the settlement founded and how long was it occupied?

3 METHODOLOGY

3.1 Survey 3.1.1 All survey work on the site was carried out using a Trimble Real Time Differential GPS survey system. All Time Team surveys, earthwork and geophysics, are compatible with each other. Surveys are related to the National Grid/ Ordnance Datum by local control using the 25” digital map. Digital copies of the survey data will be lodged with the county Sites, Monuments Record.

3 3.2 Geophysical survey 3.2.1 The instruments used were a Geoscan FM36 gradiometer, and an RM15 resistance meter. The survey areas were divided into 20x20m grids and sampled at 0.5m intervals/ 1m transects (magnetometer) and 1m interval/ 1m transects (resistance meter). For Ground Penetrating Radar, the instrument used was a Pulse Ekko 1000 instrument with various antennae, in transects, typically 1.0m/ 0.5m. The results were analysed using a mixture of GSB and commercial software.

3.3 Strategy 3.3.1 The strategy was to carry out geophysical survey in the playing fields, to determine the presence of any settlement there and to identify potential targets for excavation. Whilst this was taking place, trenches were opened in the back gardens of Arosfa and Edlyngar, where excavations were undertaken in the 1960’s, to re-examine the remains there, and record them thoroughly. Geophysical survey then took place in the car park of the Blue Bell public house and in the garden of the adjacent house (The Glen), in the back garden of the Methodist chapel and in a metalled area between garages immediately to the north of Arosfa

3.4 Excavation and Recording 3.4.1 Where access was possible a mechanical excavator (JCB or mini-digger) fitted with a toothless bucket, was used to remove the overburden from the trenches. In the one area where machine access was not possible (the back garden of Arosfa) the trench was hand excavated. All machine work was undertaken under constant archaeological supervision and ceased at the identification of significant archaeological deposits. All trenches were then cleaned by hand and archaeological deposits were excavated. All spoil arising from the excavations was scanned with a metal-detector by an experienced metal detectorist.

3.4.2 The standard Wessex Archaeology recording systems were used and all contexts and features were recorded using standard pro-forma record sheets. A record of the full extent in plan of all archaeological deposits encountered was made, usually at a scale of 1:20; sections were drawn as appropriate, at a scale of 1:10. The OD height of all principal strata and features was indicated on appropriate plans and sections. A photographic record of the investigations and individual features was also prepared. All trenches were related to the National Grid/ Ordnance Datum by local control using the 25” digital map.

3.5 Finds 3.5.1 Objects relating to human exploitation of the area that are exposed in the course of excavation were recovered or, where recovery is impracticable, recorded. All finds were recorded by context and significant objects were recorded in three dimensions. All recovered objects were retained unless they were undoubtedly of modern or recent origin.

3.6 Environmental and scientific sampling 3.6.1 The approach to environmental sampling remained flexible. The nature of the archaeological and environmental remains and their condition was not

4 known in advance, so the policy was to sample any contexts which were undisturbed and which could potentially yield material deemed on site to further the research aims of the project. All samples were processed off site by Wessex Archaeology. All artefacts were transported to the offices of Wessex Archaeology at Salisbury where they were processed and assessed.

4 RESULTS

4.1 Introduction 4.1.1 Details of individual excavated contexts and features, the full geophysical report (GSB 2005) and results of artefact and environmental sample analyses are retained in the archive. Brief context descriptions are presented in Appendix 1: Catalogue of Trench Descriptions.

4.2 Geophysical survey (Fig 2) 4.2.1 The results from both the magnetic and resistance surveys in the playing field (SAM 164a) proved disappointing. The magnetic data were severely distorted by both modern and earlier playground furniture in the form of swings, small roundabouts, climbing frames and goal posts. In addition, metal fences and buildings on the periphery of the survey grids resulted in disturbed areas. The resistance survey identified a few regions of high resistance, however, excavation proved these to be the result of a band of river sands and gravels being cut into by medieval ridge and furrow ploughing.

4.2.2 Resistance survey in part of the Scheduled garden (SAM 164b) identified a rectilinear ‘block’ of high readings close to the edge of the river; excavation revealed that the area contained gravels, with modern inclusions, that had been used to artificially raise the levels of the ground for a car port. GPR survey in the car park identified a series of reflections that in the time slices indicated the presence of a small rectilinear feature. On subsequent excavation this proved to be a concrete and rubble surface that overlay the natural clay substrata. This was cut by a modern service pipe, which provided a section through the hard surface. No Roman artefacts were noted; the feature appeared to be Victorian in date. Resistance survey in one back garden (Alyn House) proved inconclusive as did a GPR survey of a driveway / parking area immediately to the north of Arosfa.

4.3 Possible Timber Buildings (Fig 3) 4.3.1 The heavily truncated remains of a possible timber building or buildings were recorded in trenches 1 and 6 (Fig 3). The earliest phase of deposits encountered comprise the remains of a hearth (620) in trench 6 and a group of stake holes in trench 1. The second phase comprised the severely truncated remains of a clay floor and a group of associated stake-holes in Trench 1 and a small north-south gully (614) in trench 6. It is possible that these represent the very fragmentary remains of two phases of a timber building or buildings. A possible occupation deposit (104) was recorded overlying the later floor.

4.3.2 Although no coherent structure was discernible, it is assumed that the clay floors represent two phases of flooring within a timber building, or two successive timber buildings. The stake-holes may represent internal divisions

5 within the building or buildings. A further area of probable clay floor and an associated hearth were recorded at a very similar level in trench 6, with a subsequent north-south gully pre-dating the construction of the masonry wall, approximately 7m to the north. It is possible that these deposits all represent the same building or buildings, however, the form of the structure is unclear. Analysis of environmental samples taken from the occupation deposits, floors and hearth suggest a domestic function.

4.3.3 Very few datable finds were recovered from the features and deposits that represent these possible buildings, and none are closely datable within the broad period of the 2nd century. The earlier excavations that recognised this phase of occupation suggested a date of between c. AD80 and c. AD120/25 (Blockley 1989, 140), however, as the finds from which this date was derived were not securely related to the excavated stratigraphy, this is uncertain.

4.4 Masonry Structures (Fig 4) 4.4.1 The north-south aligned wall (609) in trench 6 (Fig 4) post-dated the possible timber buildings and was the only mortar-bonded wall located during the evaluation and is almost certainly a continuation of the wall found in trench E of the 1967 excavations (Blockley 1989, Fig. 3). This was associated with two pale yellow lime mortar surfaces. Only a very small area of the surface to the east of the wall (610) was visible within the trench, however, the slightly better finish of this surface suggests that this was an internal surface. The surface to the west of the wall (615) extended to the west of a small gully, and was assumed to represent an external surface.

4.4.2 The wall was constructed of unworked, or at least only very roughly trimmed local limestone and sandstone blocks in fairly regular courses, bonded with pale grey sandy lime mortar. The vertical sided foundation trench in which it was constructed cut an earlier deposit (607/611). This was a yellowish brown sandy clay loam deposit with common mortar inclusions, possibly a levelling or dump layer deposited after the demolition of the earlier timber building or buildings. Only a very small assemblage of finds was recovered from this, and only the single sherd of samian pottery recovered was datable, to the 2nd or early 3rd century.

4.4.3 Approximately 0.30m to the west of wall 609, on a parallel alignment, was a shallow gully (625), this was 0.7m wide and 0.3m deep with slightly irregular sides and a concave base. This appears to be contemporary with the wall and possible external surface and probably served a drainage function.

4.4.4 Trench 1 was positioned over one of the 1960s excavation areas (Blockley 1989,Fig. 3, trench D). The backfill of the 1960s trench was removed by hand to reveal the north-western corner of a wall (105). This wall was constructed of unworked local sandstone and limestone blocks, up to 0.55m x 0.30m x 0.25m, bonded with a light yellowish brown silty clay. The earlier excavations had removed overlying deposits, down to a yellowish brown sandy clay layer (111) to the east of the wall, and to the probable surface of a possible levelling or dump deposit (110) to the west and south of the wall, leaving the remains of the wall standing on a narrow strip of earlier deposits.

6 4.4.5 Although two small sherds of modern pottery, a fragment of modern glass and a clay pipe stem were found in association with this wall, these are thought to have been deposited during the backfilling of the earlier trench. Other finds recovered from within the fabric of the wall, comprising ceramic building material, opus signinum and pottery, including Severn Valley ware, Black Burnished ware and Whiteware, suggest that the wall was constructed in the Romano-British period, probably the 2nd century.

4.4.6 Trench 2 was excavated over a 1960s trench, with a small extension to the north excavated into a previously uninvestigated area. This located the curving wall, thought to be an apse. The wall (probably constructed in the 2nd century) is of dry stone construction and is rather irregular. The rather sinuous form of this wall suggests that it is very unlikely to represent an apse in a substantial building and is more probably some form of enclosure or compound wall. The wall appears to have fallen out of use and been substantially robbed, possibly as early as the late 2nd or early 3rd century. The only other features and deposits found in this trench comprise a layer of stones to the west, possibly formed by the collapse or demolition of the wall, and a small pit to the east.

4.4.7 Although the wall was preserved in situ and no excavation of the wall was undertaken and no dating evidence was recovered, a subsoil (202/205) appeared to have formed against the wall, probably prior to its demolition/robbing. A fairly closely datable P-profiled brooch with triangular headplate and bow consisting of three separate ribs, a late 2nd to early 3rd century type, was recovered from this deposit, along with pottery of a broadly similar date.

4.4.8 A possible alignment of masonry was visible in the gravelled surface of the car park to the north of Arosfa and trench 4 was excavated to investigate this possible feature. The masonry alignment proved to be the upper surviving course of a relatively recent brick and masonry wall, possibly part of a small shed or out-house depicted on the 1897 OS map of the area. The foundations of this wall comprised a single course of unworked local limestone and sandstone blocks bonded with a pale yellow lime mortar. It is possible that this could represent a continuation of the curvilinear wall found in trench 2 that had been utilised as the foundations of a much later building. However, as the wall exposed in trench 2 was a dry stone construction and the trench 4 foundations were mortared this appears unlikely. No datable material was recovered from these foundations and the overlying deposits appeared to be of 19th or 20th century date.

4.5 Other Features and Deposits 4.5.1 Two features were located in trench 6 that appeared to post-date the remains of the earliest possible timber building and the subsequent dump or demolition layer but to pre-date the construction of the masonry wall and associated features and deposits. These comprised a small drainage gully (617) (Fig 3) and a small scoop type feature (605). No finds were recovered from either of these features, but a 2nd century date can be inferred from their stratigraphic relationships with dated features and deposits. The function of these features is

7 uncertain due to the very limited size of the excavation area. A small, probably sub-rectangular pit (212) was recorded immediately to the east of the curving wall in trench 2 (Fig 4). The fill of this pit was sealed below the subsoil that appears to have formed prior to the demolition of the wall, but finds recovered from the fill indicate a broadly similar date.

4.5.2 Trench 9, which was located approximately 14m to the north-west of trench 6 and the substantial masonry building, contained only two features, a small ditch or gully aligned roughly north-south and a possible post-hole (Fig 4). This feature was cut into the natural substrata and sealed below the modern metalled surface, indicating that the area had probably been truncated, possibly by terracing to create the modern car park. The assemblage of finds recovered from this feature comprised mainly Romano-British ceramic building material, animal bone and fired clay fragments, along with 7 sherds of Romano-British pottery, once again broadly datable to the 2nd or early 3rd century. Environmental samples taken from the fill of this feature produced wood charcoal, but no plant macrofossils.

4.5.3 Romano-British pottery was recovered from deposits in trenches 5 and 8, which were excavated in the back garden of Alyn House in the southern end of the village. A deep deposit of probable alluvial origin, sealed below relatively recent post-medieval or modern deposits was recorded in trench 5 and a similar, though far more shallow deposit was recorded in trench 8, where it sealed a possible buried soil horizon. Both trenches were close to the present course of the Nant-y-Ffrith and it is assumed that these deposits represent seasonal overbank flooding. The possible alluvial deposit in trench 5 and the possible buried soil in trench 8 both produced small assemblages of Romano- British pottery, broadly datable to the 2nd century, but no features were located in either trench.

4.5.4 Although a small assemblage of both Romano-British and post-medieval finds were recovered from the topsoil and subsoil in trench 3, which was located in the playing field Scheduled Area (164a), no features or deposits of archaeological significance were encountered. The subsoil deposit in trench 3 may represent the remains of ridge and furrow earthworks, the very truncated remains of which were vaguely discernible within much of the rest of the field.

4.5.5 Trenches 7 and 10 were excavated in the car park of the Blue Bell public house and in the garden of the house (The Glen) immediately to the north, within the Scheduled Area (164b). Both trenches revealed only post-medieval and modern features and deposits directly overlying the natural substrata. A metalled surface, possibly associated with the cottages that once stood in this area was found in trench 7. This was cut by a modern service trench that contained a salt-glazed sewage or water pipe and directly overlay the natural substrata. The only deposits encountered in trench 10 comprised modern made-ground deposits, probably associated with a car-port which stood here in the 1970’s.

8 5 FINDS

5.1 Introduction 5.1.1 Finds were recovered from all ten of the trenches excavated although significant quantities were encountered only in Trenches 1, 2, 6 and 10. All finds have been quantified by material type within each context. Quantified data form the primary finds archive for the site, and these data are summarised by trench in Table 1. The assemblage is largely Romano-British in date, relating to the settlement underlying the present village, with a smaller quantity of post-medieval material reflecting the relatively recent foundation of the village in the 19th century.

5.1.2 Subsequent to quantification, all finds have been at least visually scanned in order to gain an overall idea of the range of types present, their condition, and their potential date range. Spot dates have been recorded for selected material types as appropriate (pottery, ceramic building material, clay pipes). All finds data are currently held on an Access database.

5.1.3 This section presents an overview of the finds assemblage, on which is based an assessment of the potential of this assemblage to contribute to an understanding of the site in its local and regional context, in particular the nature and development of the Romano-British settlement, including a consideration of potential pre-Roman origins, and the possibility of a military presence on the site after the conquest. Some comparisons can be made with material recovered from earlier, unsystematic excavations on the site (Blockley 1989).

5.2 Pottery 5.2.1 Pottery provides the primary dating evidence for the site. The largest proportion of the assemblage is of Romano-British date, with a smaller proportion of post-medieval wares. The assemblage has been quantified within each context by ware type, where known (e.g. Black Burnished ware), or by broad ware group (e.g. miscellaneous oxidised wares); totals are presented in Table 2. Spot dates have been recorded on a context by context basis, and the presence of diagnostic vessel forms noted. A high degree of residuality is apparent within the assemblage, with much of the Romano-British material (93 sherds) occurring in post-medieval contexts.

Romano-British 5.2.2 The assemblage includes a range of local, regional and imported wares indicative of a relatively high status settlement with wide-ranging contacts. Imports comprise Spanish Dressel 20 amphorae, and samian, the latter consisting almost entirely of Central Gaulish products. Samian vessel forms present include at least three decorated bowls (all, apparently, form 30), a form 18/31 platter and a form 33 cup, indicative of a 2nd to early 3rd century date range.

5.2.3 The number of mortaria could be considered unusually high for the size of assemblage. There are at least eight vessels represented here, deriving from at

9 least three production centres – north Wales or Cheshire Plain (including the ‘legionary’ kiln at Holt), Mancetter-Hartshill, and Wilderspool.

5.2.4 The coarsewares are dominated by oxidised wares, amongst which Severn Valley wares have been distinguished. Other wares potentially derive from north Wales or the Cheshire Plain (again including Holt), Wilderspool or Wroxeter. There are a few white-slipped examples. Very few diagnostic sherds are present; only bowls and everted rim jars have been identified in these wares. The coarse greywares (from unknown sources) only appear in jar forms. The presence of Black Burnished ware (BB1) from south Dorset is not surprising given the wide distribution of this ware; vessel forms present here (everted rim jar and flanged bowl) date from the later 2nd or early 3rd century AD). The single sherd of whiteware is from a flagon handle.

5.2.5 Overall, the range of fabrics and forms suggests a date range focusing on the 2nd to 3rd century AD. There are some sherds (for example, one sherd of South Gaulish samian, and the Dressel 20 amphorae) which could be earlier, but no contexts could be definitively dated earlier than 2nd century AD, and there is no evidence to suggest a pre-conquest element within the assemblage. In contrast, earlier excavations on the site produced an assemblage falling largely within the date range c.AD 80-180 (Blockley 1989, 163).

Post-Medieval 5.2.6 A large part of the post-medieval assemblage comprises sherds of modern redware flowerpots and modern refined whitewares. Other redwares, mostly glazed and sometimes white-slipped, could include some slightly earlier sherds, but the emphasis of the assemblage is definitely on the modern period, in line with the known recent origin of the village in the 19th century.

5.3 Ceramic Building Material 5.3.1 The majority of the ceramic building material recovered is of Romano-British date. This material is very fragmentary and abraded; the few diagnostic forms identified include three imbrices and four tegulae. One small fragment with an incised lattice design could be from a box flue tile. The distribution of the Romano-British ceramic building material correlates well with that of the pottery of this date and, as for the pottery, much of this material was found as residual fragments in later contexts.

5.3.2 The remaining, smaller part of the assemblage comprises fragments of modern bricks, with one wall tile.

5.4 Opus Signinum, Mortar and Wall Plaster 5.4.1 Other building materials are present in the form of opus signinum, mortar and wall plaster.

5.5 Fired Clay 5.5.1 The fired clay consists of small, abraded fragments of uncertain date and origin. One piece (from 504 –probable alluvial deposit) has been heavily heat- affected to the point of vitrification.

10 5.6 Clay Pipe 5.6.1 The clay pipes consist mostly of plain stem fragments. One complete decorated bowl, with stamped spur (?shield motifs), came from Trench 10 (buried topsoil), and one stem/spur fragment from a similar decorated bowl from Trench 2 (topsoil/backfill). Both are of 19th century or later type.

5.7 Stone 5.7.1 Of the small amount of stone recovered, none is indisputably worked or utilised. However, four fragments could represent building material – flattish sandstone ?tiles from modern overburden contexts in Trenches 1 and 2, limestone ?tessera from Trench 6 (gully 625), and a small piece of ?marble from Trench 2 (topsoil/backfill), while a possible quern fragment came from Trench 6 (layer 603).

5.8 Glass 5.8.1 All but one piece of glass is modern (vessel and window). The single piece (from Trench 2 subsoil) is from a pale blue ribbed vessel of Romano-British date.

5.9 Slag 5.9.1 The slag recovered is of uncertain date. Two pieces of ironworking slag (46g in total) came from Romano-British contexts in Trench 6 (layer 603, gully 625), but this very small quantity is insufficient to postulate on-site ironworking. The remaining slag is likely to be of relatively recent origin.

5.10 Metalwork 5.10.1 Metalwork includes objects of copper alloy, iron and lead. No coins were recovered. The copper alloy comprises a disc (probably post-medieval, from modern wall footings in Trench 6), a buckle fragment, and a tinned, sprung P- profiled brooch (Trench 2 subsoil) with triangular headplate and bow consisting of three separate ribs, a late 2nd to early 3rd century type (Böhme 1972, type 27b). This brooch type is of continental origin (although not necessarily manufacture). On the German limes sprung P-shaped brooches are regarded as ‘soldiers’ brooches’, although British evidence suggests that they were also worn by civilians (Bayley and Butcher 2004, 181). The buckle, which is of D-shaped form but which lacks either belt plate or tongue, came from a Romano-British context (pit 212). Again an object type which could be linked to a military function, the buckle is not of a particularly distinctive type and could equally well be of civilian origin.

5.10.2 Ironwork comprises nails and other probable structural items, none obviously of Romano-British date, while the lead consists of waste. The possibility that Ffrith was producing lead for, and possibly under the direction of, the Roman army stationed at nearby Chester has been noted, but lead mining in the region has a long history and these fragments could derive from later activity (all came from post-medieval contexts).

5.11 Worked Bone 5.11.1 One cuboid bone die came from Trench 2 (subsoil), of standard Romano- British form, the numerals marked with ring-and-dot motifs. The object may

11 have been fashioned from ivory, since the material is very fine grained with concentric circles

5.12 Animal Bone 5.12.1 Of the 241 bones recovered, most (231) were in fair condition with a small proportion poorly (9) or well (1) preserved. While some bones were from contexts that contained only modern or post-medieval ceramic, the majority were from Romano-British contexts (73%), with 23% from deposits containing a mixture of Romano-British and later material.

5.12.2 Gnawing incidence was in general low at 4%, and scavenger activity can not be shown to have greatly affected the assemblage. Loose teeth comprise 18% of the identified material, attesting to the fragmentary nature of some of the bones. Only 29% of bones were identified to species, also an effect of fragmentation.

5.12.3 Cattle and sheep/goat were almost equally represented, with fewer pig and small numbers of horse, dog or fox and wild animals (represented by a mouse jaw and hare radius). Bird bones were all similar in morphology to domestic fowl.

5.12.4 6% of bones could be measured and 13% aged, including several bones from young pigs, and sheep that died at a point before the lower third molar began to be worn. Juvenile cattle remains were recovered from 603. A cleft in the proximal articular surface of a cattle metacarpal is probably a congenital abnormality.

5.12.5 Butchery marks were observed on 17% of bones, although many of these were probably modern saw marks. Common marks were chops for portioning and to a lesser extent cuts from disarticulating and filleting meat from joints. One bone in 603 was ivoried, and several contexts contained sheep and cattle sized ribs that had been chopped into sections. 13% of bones had been burnt, most from 618. A cattle metatarsal in 614 with extensive discolouring and cracking on the shaft may also have been heated, perhaps to more easily extract the marrow.

5.12.6 No unusual combinations of bone elements, that might represent particular activities, were noted.

6 PALAEOENVIRONMENTAL EVIDENCE

6.1 Five bulk samples were processed by standard flotation methods and the results quantified in Table 3. The flots had little or no roots, and so the material can be regarded as reasonably well-sealed with the likelihood of little intrusive material.

6.2 There were relatively few plant remains within the samples and only one sample, from hearth 620, contained any cereal remains. One sample from the occupation deposit (context 104) contained only a single seed of cleavers

12 (Galium aparine), that from context 110 contained four fragments of onion couch grass tuber (Arrhenatherum elatius ssp. bulbosum) as well as 22 fragments of root stem.

6.3 Onion couch grass grows high and can form thick swathes of dry grassland in the summer. It is thought to have been used as tinder, and is probably cleared, by uprooting, to form a firebreak, only the dry stems with some tubers attached then being used as tinder. Their presence within the floor deposit along with reasonable amounts of wood charcoal may indicate their use for fire lighting, although no such remains were found within the hearth deposit, hearth 620, context 618.

6.4 In addition this sample also contained two fragments of hazelnut, three inner seed kernels probably from buttercup (Ranunculus acris/repens/bulbosus), and a seed of ribwort plantain (Plantago lanceolata). The sample also contained a single seed of possible false-oat grass, or a species with similar sized seeds. It is probable that the seeds of buttercup and plantain arrived with the uprooted onion couch grass.

6.5 The sample from the possible hearth (620) was the only one that contained cereal remains. These were some 10 grains of hulled barley (Hordeum vulgare sl). A single grain of garden pea (Pisium sativum) was also recovered along with some 20 fragments of hazelnut. Only a single seed of persicaria (Persicaria maculosa/ lapathifolia) could be classified as a possible weed. Of all the deposits examined only this would seem to represent what might be expected from a domestic context.

6.6 The remaining two samples from gully 901 had no identifiable plant macrosfossils, containing only wood charcoal.

6.7 Charcoal was noted from the flots. All the samples contained a few to several larger fragments of wood charcoal. The sample from the hearth 620 contained numerous large fragments of probable heart wood. Those from the gully also contained more fragmented charcoal, although that from sample 5, context 902, had a few larger fragments.

6.8 Only very small numbers of shells of land snails were recovered in the samples. That from the occupational floor had a few shells of Helicella itala, Cochlicopa spp. and Vallonia spp. That from the hearth had single shells of Vallonia spp. and Discus rotundatus and three of Aegopinella sp.

6.9 Several of the flots contained fragments of animal bone. Small numbers of fish bone were recovered from the samples, a few scales from the floor deposit and single finds of vertebrae from the hearth and gully deposits. The hearth sample also contained several fragments of burnt bone, as well as tail bones from a small mammal.

13 7 DISCUSSION

7.1 This project attempted to resolve several questions about the origins, date, longevity, nature, function and extent of the Romano-British site at Ffrith (see section 2.2, above). The results enable some of these questions to be addressed in more detail than previously, however, they do not provide definitive answers and some of the research questions remain unresolved. The evaluation broadly confirmed the previous dating and longevity of the settlement (Blockley 1989) but provided no clear indication of its function, nature or economic basis. Although Romano-British material was recovered from a wide area within the village the majority of this material appears to represent residual finds within later deposits. The only features and deposits of undoubted Romano-British date were all found within a small area in the vicinity of the 1960s excavations, which appears to represent the main focus of the settlement. It is perhaps significant that geophysical survey and limited trial trenching failed to identify any Romano-British features or deposits within either of the scheduled areas.

7.2 The remains to two phases of clay flooring were recognised during these and the earlier excavations. The form and function of this building or buildings is uncertain, but environmental remains recovered from the hearth suggest that a domestic function is likely. Only very small areas of the floors and associated features were available for investigation and very few diagnostic finds were recovered. The finds assemblages from these deposits can only be broadly dated to the 2nd century, however, when their stratigraphic relationship to the similarly dated masonry buildings is considered, an early 2nd century date is indicated.

7.3 The dry stone, clay bonded and mortared walls all appear to be broadly contemporaneous. However, the very different styles of construction and degree of finish suggests that these probably represent three separate structures. The mortared wall (609) in trench 6 was the only mortared wall found and appears to represent a substantial, well-constructed building. Earlier archaeological work suggests that at least one room in this building contained a hypocaust, which, along with some of the finds, indicates a relatively high status. The character and function of the structure represented by the clay bonded wall in trench 1 (105) is uncertain, however, its fairly crude construction suggests that it is unlikely that this was part of the substantial building represented by the mortared wall in trench 6. The similarity of alignment between the two walls indicates that they may be associated and the clay bonded wall may represent a small ancillary building to the south, possibly a workshop or storage building. The curvilinear wall located in trench 2 (206/7), was of dry stone construction and rather irregular. It is clear that this does not represent an apse ended room, indeed, its rather crude construction and irregular form suggest that this is more likely to represent some form of enclosure wall, probably enclosing the building or buildings represented by the other walls. Although residual Romano-British finds were recovered from all of the trenches in the south of the village no features or deposits indicative of settlement were encountered in this area, suggesting that the settlement was

14 centred around the buildings in the north, and may even have comprised just a single substantial building and ancillary structures within a walled enclosure.

7.4 The small finds assemblage has served to date the Romano-British activity within the excavated areas to the 2nd/3rd centuries AD, in contrast to the slightly earlier date range (late 1st to late 2nd century AD) suggested for material from the earlier excavations (Blockley 1989). However, some material from the present investigations could indeed be slightly earlier, indicating that the earlier timber buildings may have been constructed as early as the late 1st century. The settlement may have continued into the 3rd century, but very little material dated later than the early 3rd century has been found in Ffrith, suggesting that the settlement was abandoned, or continued in a much reduced form, after this date.

7.5 Evidence for on-site metalworking (lead, iron) is ambiguous, as is the possibility of a military presence on the site. The buckle, the brooch associated with military use in Germany (but not necessarily in Britain), and the mortaria and other pottery possibly supplied by the military depot at Holt, are hardly conclusive evidence in this respect, although they can be added to a handful of items of military equipment found during the earlier excavations, and an association with the Twentieth Legion as evidenced by two stamped ceramic tiles (ibid.). However, given the proximity of the legionary fort at Chester and the Holt depot, it is perhaps likely that these represent the nearest source of ceramics and do not necessarily imply a military function for the settlement.

7.6 No clear indication of the function of the settlement has been recovered by any of the excavations. The range of finds is fairly limited, and there is little that can be used to characterise the nature of the settlement beyond indicating the presence of substantial buildings (ceramic tiles, opus signinum, wall plaster) and inhabitants with access to fineware pottery and glass. Presumably local resources, such as lead ore, limestone and coal, were being exploited, but which ones is open to question. Lead ore is present in the local area and may have been exploited, but the local seams appear to be quite thin and would not have produced much (M. Walters pers. comm.) so this is perhaps unlikely. The other notable local resource is the limestone, which has been exploited recently and may have been during the Romano-British period. The local coal measures could have provided the fuel for lead smelting or lime roasting to produce mortar.

8 RECOMMENDATIONS

8.1 The majority of the Romano-British features and deposits identified in the evaluation were located in the vicinity of the 1960s excavations and many had been previously exposed and recorded. Further analysis of these would not enable any of the unanswered or partly answered research questions to be addressed in more detail and no further work is suggested.

8.2 The potential of the finds assemblage to further elucidate the nature of the Romano-British settlement is limited by its small size, relatively restricted range, and high degree of residuality. This assemblage alone would not

15 warrant further detailed analysis, although it would form a useful addition to any overall assemblage from previous or future fieldwork on the site.

8.3 Some selective discard of material can be suggested, comprising the post- medieval component within the assemblage (pottery, brick/tile, clay pipe, glass), as well as undiagnostic material (slag, stone, iron, animal bone) from post-medieval contexts.

8.4 The environmental remains recovered comprise charred plant remains, charcoal, land snails and small mammal and fish remains. The charred plant remains indicate the possible presence of more domestic activities within the hearth sample and hence the survival of material relating to such activities, but have no further potential. Charcoal was recovered from several of the samples and this has the potential to reveal the nature of the utilisation of woodland species and possibly the deliberate selection of certain species for fuel upon the site. However, only the charcoal recovered from the hearth was potentially burnt in situ and no further analysis is recommended. While land snail remains were present they were only recovered in very low numbers, and none of the assemblages are either out of the ordinary, nor would help with the specific site questions, and therefore are considered to have no further potential. The fish bones and scales recovered from the occupation deposits overlying the second phase of clay floors are probably a part of food remains. The identification of these would help determine the fish eaten and possibly the use of the room excavated, however, as the form and even the size of the structure the floor represents is uncertain no further work is suggested.

8.5 Given the above assessment of the results of the evaluation, no further analysis of the structural, finds or environmental data is considered to be necessary. A copy of this report will be submitted to the Flintshire Sites and Monuments Record. It is recommended that a brief report summarising the evaluation and incorporating the results of this assessment be published in the Flintshire Historical Society Journal along with a similar note in the annual summary of work on Romano-British archaeology in Britannia (Roman Britain in 2005).

9 ARCHIVE

9.1 The archive, which includes all artefacts, written, drawn and photographic records relating directly to the investigation is undertaken, is currently held at the offices of Wessex archaeology under the site code FFR 05 and Wessex Archaeology project no. 59461. The paper archive is contained in one lever arch file. In due course, Time Team will transfer ownership of the archive to County Museum Service.

16 10 REFERENCES

Bayley, J. and Butcher, S., 2004, Roman Brooches in Britain: a technological and typological study based on the Richborough Collection, Rep. Res. Comm. Soc. Antiq. London 68

Blockley, K., 1989, Excavations on the Romano-British Settlement at Ffrith, Clwyd, 1967-9. Flintshire Historical Society Journal, vol. 32, 135-165.

Böhme, A., 1972, ‘Die Fibeln der Kastelle Saalburg und Zugmantel’, Saalburg Jahrbuch 29, 5-112

Davies, E., 1949, The Prehistoric and Roman Remains of Flintshire.

GSB 2005, Geophysical Survey Report 2005/21: Ffrith, Flintshire, Wales. GSB Prospection Ltd, unpublished client report.

Videotext Communications 2005, Proposed Archaeological Evaluation at Ffrith, Flintshire, Wales. Unpublished project design

17 Appendix 1: Catalogue of Trench Descriptions

Trench 1 Location: Back Garden of Arosfa Max. Depth: 1.22m Dimensions: 3.30m x 1.90m Ground level-122.80m OD Context No. Description Depth Finds 101 Modern overburden (topsoil, patio etc.) and backfill 0f 1967 0-0.95m 102 trench 103 107 108 123 105 Masonry wall footing in foundation trench. Large, 0.60-0.80m unworked, local limestone blocks cemented with yellowish brown silty clay, probably forms the north-west corner of a building or structure. 122 Construction cut for wall 105. Largely removed by earlier 0.65-0.80m excavations. 104 Mid brownish grey sandy silt deposit with sparse burnt clay 0.95-1.01m and charcoal inclusions. Possible occupation deposit above floor 106 112/113 Stakeholes and fills. All cut possible clay floor 106 1.01-1.18m 114/115 116/117 118/119 120/121 106 Pale yellow silty clay with burnt clay, charcoal and mortar 1.01-1.08m inclusions. Possible clay floor below occupation deposit 104. 109 Light yellowish brown silt clay. Possible clay floor. 1.01-1.11m Probably the same as 106. 110 Mid greyish brown silty clay deposit below floors 109 and 1.08-1.27m 106. Possible levelling deposit. 124/125 Unexcavated stakeholes and fills. All seen cutting possible 1.27m+ 126/127 floor or natural substrata 111, but may be associated with 128/129 stakeholes 112-121, which were in a less truncated area. 130/131 132/133 134/135 136/137 138/139 140/141 142/143 144/145 146/147 111 Light yellowish brown sandy clay. Unexcavated. Cut by 1.27m+ several stakeholes. Possible clay floor or natural alluvial substrata utilised as a floor.

18 Trench 2 Location: Back Garden of Edlynger Max. Depth: 1.05m Dimensions: 7.65m x 3.85m Ground level-122.91m OD Context No. Description Depth Finds 201 Dark greyish brown silty clay loam topsoil and backfill of of 0-0.90m 1967 trench. 213 Modern intrusion 0.32-0.75m 214 Fill of 213 0.32-0.75m 203 Mid greyish brown silty clay loam subsoil. 0.32-0.52m 205 208 202 Mid-light greyish brown sandy clay loam subsoil with 0.52-0.70m 204 abundant limestone inclusions. 206 Curving limestone wall, survives to a maximum height of 0.52-0.85m 0.20m (two courses) and up to 0.70m wide. Aligned 207 approximately north-south, may continue in trench 4. Drystone construction with small butresses on eastern side. 211 Probably represents some sort of field or enclosure boundary rather than part of a building. 210 Mid-light greyish brown silty clay fill of pit 212 0.70-1.10m 212 Probable sub-rectangular pit to east of wall 206/207. Over 0.70-1.10m 1.25m long, 0.50m+ wide and 0.40m deep with moderately steep straight sides and a concave base. 209 Light brown sandy clay with sparse limestone inclusions. 0.70-1.10m+ Natural alluvial substrata.

Trench 3 Location: Playing Field Max. Depth: 1.10m Dimensions: 7.00m x 2.00m Ground level-120.30m OD Context No. Description Depth Finds 300 Dark greyish brown silty clay loam topsoil with common 0-0.28m limestone inclusions. 301 Mid greyish brown sandy clay subsoil with abundant small 0.28-0.37m limestone inclusions. 302 Pale-mid yellowish brown sandy clay with abundant sub- 0.37-1.10m+ rounded limestone inclusions.

Trench 4 Location: Car Park North of Arosfa Max. Depth: 0.55m Dimensions: 3.60m x 2.50m Ground level-122.32m OD Context No. Description Depth Finds 401 Modern metalled surface and make-up/bedding layer. 0-0.20m 402 Mid yellowish brown silty clay deposit with common 0.20-0.25m mortar inclusions. Probably a levelling/made-ground layer associated with the construction of a building depicted on the 1897 edition OS map (not present on 1872 version), probably a small outhouse or shed. 403 Mortar and brick fragment foundation/footing associated 0.20m-0.30m with the construction of a building depicted on the 1897 edition OS map (not present on 1872 version), probably a small outhouse or shed. 404 Brick surface, associated with building depicted on the 0.20-0.28m 1897 edition OS map (not present on 1872 version), probably a small outhouse or shed. Overlies 402. 405 Rough limestone wall bonded with pale grey lime mortar. 0-0.28m Associated with building depicted on the 1897 edition OS map (not present on 1872 version), probably a small outhouse or shed. Overlies 402. 406 Very dark grey sandy silt deposit. Pre-dates 402-5 and seals 0.28-0.32m wall 407. Possible remnant of buried topsoil.

19 407 Very short length of rough limestone wall, possible 0.32-0.55m continuation of wall 206/207 in trench 2. 408 Mid yellowish brown silty clay loam deposit. Cut by wall 0.32-0.55m 407. Possible buried topsoil. 409 Pale yellowish brown sandy clay with common limestone 0.55m+ inclusions. Natural alluvial substrata. Heavily iron-panned in places.

Trench 5 Location: Behind Alyn House Max. Depth: 2.05m Dimensions: 2.50m x 1.10m Ground level-121.98m OD Context No. Description Depth Finds 500 Dark greyish brown silty clay loam topsoil with common 0-0.35m limestone inclusions. 501 Pale greyish brown sandy clay with c. 80% limestone 0.35-0.44m rubble inclusions. Possible crude metalled surface. 502 Dark greyish brown silty clay loam, possible buried topsoil. 0.44-0.62m Modern finds noted. 504 Mid-light yellowish brown sandy clay loam subsoil. 0.62-2.00m Amphora fragments recovered from c. 200mm above base of deposit (1.50m below present ground surface). Looks like colluvium, but is perhaps more likely to be alluvial or a made-ground deposit. 503 Pale–mid yellowish brown sandy clay with abundant 2.00-2.05m+ limestone inclusions. Natural substrata.

Trench 6 Location: Garage Area North of Arosfa Max. Depth: 1.01m Dimensions: 4.20m x 3.20m Ground level-122.24m OD Context No. Description Depth Finds 601 Modern overburden. Metalled surface and bedding layer 0-0.70m 602 Modern Brick wall footings 0.25-0.40m 606 Dark greyish brown clay loam with abundant mortar, 0.55-0.68m charcoal and small pebble inclusions. Upper fill of gully 625. 608 Dark yellowish brown clay loam, basal fill of gully 625. 0.55-0.85m 625 Gully aligned approximately north-south, traced for 2.90m 0.55-0.85m across trench, 0.70m wide and 0.30m deep with irregular sides and a concave base. 604 Dark yellowish brown clay loam fill of shallow pit or scoop 0.55-0.69m 605. 605 Shallow oval pit or scoop, approximately 0.60m in diameter 0.55-0.69m and 0.09m deep with steep, concave sides and a flat base. 603 Dark greyish brown sandy loam, cut by foundations of 0.50-0.60m brick wall 602 and feature 605. 610 Very pale yellow-off white lime mortar floor to the east of 0.67-0.77m wall 609. 615 Very pale yellow-off white lime mortar surface to the west 0.55-0.67m of wall 609. 626 Very dark greyish brown silty clay loam, backfill of 0.65-0.90m+ foundation trench 627. 609 Limestone wall aligned approximately north-south. Bonded 0.40-1.00m+ with pale grey lime mortar, 0.55m wide and traced for 2.60m within the trench. Constructed within foundation trench 627. 627 Vertical sided foundation trench for wall 609, only 0.65-0.90m+ observed on west side of wall, not fully excavated. Cuts deposit 607, sealed below mortar surface 615. 628 Mid yellowish brown silty clay deposit, below 603, seals 0.60-0.75m fill of gully 617. 616 Mid greyish brown sandy loam with c. 90% small gravel 0.75-0.97m

20 inclusions. Single fill of gully 617. Sealed below deposit 628. 617 Small gully aligned approximately north-south, parallel to 0.75-0.97m wall 609, 3.00m to the east. Approximately 0.50m wide and 0.23m deep with steep, irregular sides and a flat base. Probable drainage gully or garden feature. 607 Mid yellowish brown sandy clay loam with common mortar 0.65-0.90m inclusions. Confined to west side of wall 609. Probably the same as 611 to the east of the wall. 611 Mid yellowish brown sandy clay loam to east of wall 609. 0.77-0.90m 612 Probable bedding/make-up layer for floor 610. 622 Thin layer of pea grit below deposit 603 and above 623 0.60-0.63m 623 Dark greyish brown sandy clay loam deposit. Sealed below 0.63-0.70m 622 and above 624. 624 Mid-dark yellowish brown sandy clay loam with common 0.70-0.80m+ mortar inclusions. Confined to west of gully 617, probably the same as 607 to the east of gully. Not fully excavated. 613 Mid yellowish brown sandy clay loam deposit. Predates 0.85-1.00m construction of wall 609 and associated floors. Probably the same as 614 to the west of the wall. 614 Mid yellowish brown sandy clay loam deposit. Predates 0.70-0.90m construction of wall 609 and associated floors. Probably the same as 613 to the west of the wall. 618 Very dark grey silty clay loam deposit with very abundant 0.90-0.94m charcoal inclusions. Upper fill of hearth 620 619 Mid reddish brown heat effected sandy clay. Basal fill of 0.90-1.01m hearth 620. 620 Sub-circular scoop, 0.90m in diameter and 0.11m deep with 0.90-1.01m shallow concave sides and concave base. Probable hearth. 621 Pale yellowish brown sandy clay deposit, unexcavated. Cut 1.01m+ by hearth 620. Possible clay floor or natural substrata utilised as a floor.

Trench 7 Location: Bluebell Car Park Max. Depth: 0.45m Dimensions: 3.5m x 2.4m Ground level-120.40m OD Context No. Description Depth Finds 700 Modern Overburden – Tarmac and hardcore 0-0.35m 701 North-south aligned trench – modern sewer. 0.35-0.55m 702 Fill of service trench 701 0.35-0.55m 703 Mid-light grey sandy clay with c. 80% limestone inclusions. 0.35-0.45m Probably earlier metalled surface. 704 Pale yellowish brown sandy clay with common limestone 0.45m+ inclusions. Natural Substrata.

Trench 8 Location: Behind Chapel/Alyn House Max. Depth: -1.30m Dimensions: 3.5m x 3.1m Ground level-121.96m OD Context No. Description Depth Finds 800 Dark greyish brown silty clay loam topsoil with common 0-0.20m limestone inclusions. 801 Pale greyish brown sandy clay with c. 80% limestone 0.20-0.30m rubble inclusions. Possible crude metalled surface. 802 Dark greyish brown silty clay loam, possible buried topsoil. 0.30-0.60m Modern finds recovered. 803 Light greyish brown sandy clay deposit. Looks like 0.60-0.94m colluvium, but is perhaps more likely to be alluvial or a made-ground deposit. 804 Pale yellowish brown sandy clay with frequent root marks. 0.94-1.30m Possible buried topsoil. 805 Pale – mid yellowish brown sandy clay with abundant 1.30m+

21 limestone inclusions. Natural substrata.

Trench 9 Location: Car Park North of Arosfa Max. Depth: 0.50m Dimensions: Ground level-121.98m OD Context No. Description Depth Finds 900 Modern overburden. Gravelled surface and bedding layer. 0-0.27m 901 North – south aligned gully, on average 0.80m wide and 0.27-0.50m 0.20m deep with moderately steep, concave sides and a concave base. 902 Dark brownish grey silty clay fill of gully 901. Contains 0.27-0.50m sparse charcoal flecks and common small pebbles. 903/904 Post-hole (903), 0.28m in diameter and 0.15m deep with a 0.27-0.42m pointed base. Filled with a mid brown silty clay fill with sparse charcoal flecks. 905 Mid yellowish brown sandy clay, natural. alluvial substrata. 0.27-0.50m+

Trench 10 Location: The Glen Max. Depth: 2.03m Dimensions: 2.2m x 1.2m Ground level-120.04m OD Context No. Description Depth Finds 1000 Dark greyish brown silty clay loam topsoil 0-0.20m 1001 Modern scalpings 0.20-0.80m 1002 Dark greyish brown silty clay loam. Probable recently 0.80-1.10m buried topsoil 1003 Modern cinder and coal deposit 1.10-1.20m 1004 Mid-light greyish brown sandy clay loam sub-soil. 1.20-2.03m Becomes gradually darker with depth. Sparse charcoal inclusions. 1005 Mid-light yellowish brown sandy clay with common 2.03m+ limestone inclusions. Natural substrata.

22 Table 1: Finds totals by trench (number / weight in grammes)

Material Tr 1 Tr 2 Tr 3 Tr 4 Tr 5 Tr 6 Tr 7 Tr 8 Tr 9 Tr 10 TOTAL Pottery 64/500 139/1764 1/7 3/10 7/273 28/713 - 8/296 7/331 49/228 306/4122 Romano-British 40/247 111/1485 1/7 1/1 6/265 23/500 - 4/204 7/331 - 193/3040 Post-Medieval 24/253 28/279 - 2/9 1/8 5/213 - 4/92 - 49/228 113/1082 Ceramic Building Material 66/2738 98/5261 4/212 5/246 5/67 17/683 - - 10/379 3/156 208/9742 Romano-British 53/2451 82/4939 2/18 3/21 5/67 17/683 - - 10/379 - 172/8558 Post-Medieval 13/287 16/322 2/194 2/225 - - - - - 3/156 36/1184 Opus Signinum 1/73 - - - - 1/13 - - - 2/86 Mortar 1/829 - - - - 2/116 - - - 1/158 4/1103 Wall Plaster 2/21 - - - - 10 - - - 12 Fired Clay - 3/36 - - 1/5 4/56 - - 10/89 3/14 21/200 Clay Pipe 3/4 4/11 - - - - 2/2 - - 1/9 10/26 Stone 5/163 5/71 - - - 9/646 - - - 4/75 23/955 Glass 5/98 4/13 - - - - - 3/12 1/6 14/180 27/309 Slag 2/48 - - - - 2/46 9/634 - - 1/123 14/851 Metalwork 6 3 2 1 - 3 2 - - - 17 Copper Alloy - 2 - - - 1 - - - - 3 Iron 5 - - 1 - 2 2 - - - 10 Lead 1 1 2 ------4 Worked Bone - 1/5 ------1/5 Animal Bone 38/662 86/810 - 15/89 - 114/434 2/24 - 15/40 3/27 273/2086 Table 2: pottery totals by ware type

Date Range Ware type No. sherds Weight (g) ROMANO-BRITISH Samian 13 151 Amphora 4 493 Mancetter-Hartshill mortarium 1 52 South Wales mortarium 12 554 Wilderspool mortarium 4 469 Greyware 46 393 Severn Valley ware 45 227 Misc. oxidised wares 53 560 Black Burnished ware 14 127 Whiteware 1 14 sub-total Romano-British 193 3040 POST-MEDIEVAL Redware 46 603 Stoneware 8 56 Refined redware 4 9 Bone china 6 33 Refined whiteware 47 280 Sanitary ware 2 101 sub-total post-medieval 113 1082 OVERALL TOTAL 306 4122

Table 3. Assessment of the charred plant remains and charcoal.

Flot Residue Feature type/ Context Sample size flot size Grain Chaff Weed Other Charcoal Other Charcoal no litres ml uncharred charred >5.6mm >5.6mm Romano-British (Prob. early) Occupation layer/floors 104 1 10 80 5 - - - C B moll-t (C) - fish (C) smb (C) 110 3 22 40 5 - - c B(h) B moll-t (C) - Hearth 620 618 2 12 490 - A - - A(h) A* moll-(C) - fish (C) smb (C) Gully 901 902 4 13 45 5 - - - - A fish - (C) - 901 902 5 15 40 5 - - - - B smb-(C)

KEY: A** = exceptional, A* = 30+ items, A = t10 items, B = 9 - 5 items, C = < 5 items, (h) = hazelnuts, smb = small mammal bones; Moll-t = terrestrial molluscs Moll-f = freshwater molluscs; Analysis, C = charcoal, P = plant, M = molluscs NOTE: 1flot is total, but flot in superscript = ml of rooty material. 2Unburnt seed in lower case to distinguish from charred remains

Geophysical data courtesy of GSB Prospection Ltd Trench 6 - Phase 1 Phase 2

620 614

0 1 2m

Trench 1 - Phase 1 Phase 2

106

109

0 1 2m

Modern brick wall Stone

Unexcavated Stakehole This material is for client report only © Wessex Archaeology. No unauthorised reproduction. Date: 20/06/05 Revision Number: 0 Modern intrusion Clay floor Wessex Scale: 1:50 @ A3 Illustrator: MSM Archaeology Path: Y:\PROJECTS\59461\D.. O..\Report Figures (05-06)\Time Team\05-06-16

Trench 1 & 6 Plans showing phasing of features Figure 3

THE TRUST FOR WESSEX ARCHAEOLOGY LTD. Head Office: Portway House, Old Sarum Park, Salisbury, Wiltshire SP4 6EB. Tel: 01722 326867 Fax: 01722 337562 [email protected] www.wessexarch.co.uk London Office: Unit 701, The Chandlery, 50 Westminster Bridge Road, London SE1 7QY. Tel: 020 7953 7494 Fax: 020 7953 7499 [email protected] www.wessexarch.co.uk

Registered Charity No. 287786. A company with limited liability registered in England No. 1712772.